Computational imaging and occluded objects perception method based on polarization camera array.

Traditional optical imaging relies on light intensity information from light reflected or transmitted by an object, while polarization imaging utilizes polarization information of light. Camera array imaging is a potent computational imaging technique that enables computational imaging at any depth. However, conventional imaging methods mainly focus on removing occlusions in the foreground and targeting, with limited attention to imaging and analyzing polarization characteristics at specific depths. Conventional camera arrays cannot be used for polarization layered computational imaging. Thus, to study polarization layered imaging at various depths, we devised a flexible polarization camera array system and proposed a depth-parallax relationship model to achieve computational imaging of polarization arrays and polarization information reconstruction under varying conditions and depths. A series of experiments were conducted under diverse occlusion environments. We analyzed the distinctive characteristics of the imaging results obtained from the polarization array, employing a range of array distribution methods, materials, occlusion density, and depths. Our research successfully achieved computational imaging that incorporates a layered perception of objects. Finally, we evaluated the object region's polarization information using the gray level co-occurrence matrix feature method.

Design of custom-made navigational template of femoral head and pilot research in total hip resurfacing arthroplasty.

BACKGROUND: To develop a novel custom-made navigational template for accurate prosthesis implantation in total hip resurfacing arthroplasty (THRA) by computer-aided technology. METHODS: The template was produced based on data preoperatively acquired from computed tomography (CT) scan. The position of the drill guide was obtained according to the anatomical axis of the femoral neck which was defined by the point of the femoral head center and another point of the femoral neck center. The final direction of the drill guide was confirmed by a valgus angle. The surface of the template was constructed based on the inverse of the femoral neck surface. Then the template was made of acrylate resin by using rapid prototyping (RP) technique. Finally, all the templates were verified in 17 cadavers arranged for THRA and postoperative medical images were employed to evaluate the accuracy and validity of the template. RESULTS: The templates had achieved a high fitting with the femoral neck surface, and there were no guide failures. Postoperative evaluation revealed that the Kirschner-wires pass through the center of the femoral head and femoral neck, presenting a relative expected and acceptable valgus angle to the central axis of the femoral neck. The lateral offset showed the relative valgus angle achieved as expected, the horizontal offset showed that no obvious antero-posterior deviation occured. The comparison between the preoperative Neck-shaft angle (NSA) and the postoperative Stem-shaft angle (SSA) showed there is no significant difference(P > 0.05). CONCLUSION: The novel custom-made navigational template of femoral head can effectively assist surgeons for accurately implanting the femoral head components to the desired position in THRA.

Inverse solutions for a Risley prism scanner with iterative refinement by a forward solution.

Risley prism scanners are increasingly used for laser beam steering due to their wide angular scanning range and high resolution. However, the inverse problem, which focuses on obtaining the required prisms' orientations for a given target position, has not been perfectly solved so far. The existing inverse solutions are not accurate or efficient enough for high-accuracy and real-time tracking. An iterative method that combines an approximate inverse solution with an iterative refinement by the forward solution is set forth in this paper. Two case studies indicate that the rotation motions of Risley prism pairs controlled by iterative solutions can slew the beam to create the desired tracking pattern quickly and accurately. Based on this method, a Risley prism scanner developed as a standard trajectory generator is implemented for the error measurement of a robotic manipulator in our experiments. The simulation and experimental results show that the inverse solution for one target point can be obtained within nine iterations for a prescribed tracking error threshold.

Macroscopic-and-Microscopic Rain Streaks Disentanglement Network for Single-Image Deraining.

Single-image deraining aims to restore the image that is degraded by the rain streaks, where the long-standing bottleneck lies in how to disentangle the rain streaks from the given rainy image. Despite the progress made by substantial existing works, several crucial questions - e.g., How to distinguish rain streaks and clean image, while how to disentangle rain streaks from low-frequency pixels, and further prevent the blurry edges - have not been well investigated. In this paper, we attempt to solve all of them under one roof. Our observation is that rain streaks are bright stripes with higher pixel values that are evenly distributed in each color channel of the rainy image, while the disentanglement of the high-frequency rain streaks is equivalent to decreasing the standard deviation of the pixel distribution for the rainy image. To this end, we propose a self-supervised rain streaks learning network to characterize the similar pixel distribution of the rain streaks from a macroscopic viewpoint over various low-frequency pixels of gray-scale rainy images, coupling with a supervised rain streaks learning network to explore the specific pixel distribution of the rain streaks from a microscopic viewpoint between each paired rainy and clean images. Building on this, a self-attentive adversarial restoration network comes up to prevent the further blurry edges. These networks compose an end-to-end Macroscopic-and-Microscopic Rain Streaks Disentanglement Network, named [Formula: see text]RSD-Net, to learn rain streaks, which is further removed for single image deraining. The experimental results validate its advantages on deraining benchmarks against the state-of-the-arts. The code is available at: https://github.com/xinjiangaohfut/MMRSD-Net.

An Interpretable Deep Architecture for Similarity Learning Built Upon Hierarchical Concepts.

In general, development of adequately complex mathematical models, such as deep neural networks, can be an effective way to improve the accuracy of learning models. However, this is achieved at the cost of reduced post-hoc model interpretability, because what is learned by the model can become less intelligible and tractable to humans as the model complexity increases. In this paper, we target a similarity learning task in the context of image retrieval, with a focus on the model interpretability issue. An effective similarity neural network (SNN) is proposed to offer not only to seek robust retrieval performance but also to achieve satisfactory post-hoc interpretability. The network is designed by linking the neuron architecture with the organization of a concept tree and by formulating neuron operations to pass similarity information between concepts. Various ways of understanding and visualizing what is learned by the SNN neurons are proposed. We also exhaustively evaluate the proposed approach using a number of relevant datasets against a number of state-of-the-art approaches to demonstrate the effectiveness of the proposed network. Our results show that the proposed approach can offer superior performance when compared against state-of-the-art approaches. Neuron visualization results are demonstrated to support the understanding of the trained neurons.

Cooperation of HDL Receptor and Hepatic Lipase in the Selective Uptake of HDL(2)-CE by Rat Hepatic Sinusoidal Cells.

The recombined (3)H-CE-HDL(2)(rHDL(2)) keeps the biological activities of the native HDL(2). After rat hepatic sinusoidal cells were incubated with rHDL(2) at 37 degrees for 3 h (normal group), the cell-endocytic cpm was 995-/+147(mean-/+s, n=2). After the cells were further incubated for 2 h, the cell-release TCA-precipitable cpm and the TCA-supernatant cpm were 78-/+32 and 12-/+9 respectively. These values were 339-/+62, 19-/+11 and 9-/+5 respectively in the acetylimidazole-modified group, and 542-/+78, 34-/+14 and 9-/+8 respectively in the heparin-pretreated group. Our results suggested that: (1) Rat hepatic sinusoidal cells internalize HDL(2) and take up HDL(2)-CE by its HDL receptor, and HDL(3) was secreted out of the cells by retroendocytosis. (2) Hepatic lipase (HL) induces directly the selective uptake of HDL(2)-CE by the cells. (3) There is cooperation between the HDL receptor and HL in the selective uptake of HDL(2)-CE by the cells.